ChapterPDF Available

The better-than-average effect

The Better-Than-Average
Ohio University
Ohio State University
Most people are average but few people believe it.  e tendency to
evaluate oneself more favorably than an average peer is one of social
psychology’s chestnuts—a fi nding that will never let you down when
running a class demonstration.  is better-than-average eff ect has been obtained
in numerous studies, with diverse populations, on multiple dimensions, and with
various measurements techniques.
e better-than-average eff ect is a particular type of social comparison, one in
which people compare their characteristics or behaviors against a norm or standard,
which is usually the average standing of their peers on the characteristic. In this
regard, the better-than-average eff ect falls outside the mainstream of traditional
social comparison theory. Following Festinger (1954), social comparison theorists
have emphasized the precursors and consequences of comparisons between people.
Arguably, however, comparisons with normative standards are at least as prevalent
as interpersonal comparisons.  e self versus average peer judgments studied in
better-than-average eff ect research are akin to social comparisons such as assessing
whether one is meeting a group’s moral standards or performance expectations.
e better-than-average eff ect is considered to be one of the most robust of
all self-enhancement phenomena (Taylor & Brown, 1988; Sedikides & Gregg,
2003).  e better-than-average eff ect shares this distinction with the optimistic
bias—the tendency to overestimate one’s chances of good fortune and to under-
estimate one’s risk for misfortune. Whereas the better-than-average eff ect pertains
to self versus average peer comparisons on behavior and trait dimensions, the
optimistic bias involves comparisons about life events such as winning the lottery
or getting divorced. Although we concentrate on the better-than-average eff ect in
Alicke_RT19523_C05.indd 85Alicke_RT19523_C05.indd 85 7/11/2005 11:06:44 AM7/11/2005 11:06:44 AM
this chapter, many of the issues underlying better-than-average judgments apply
as well to relative risk assessments. Connections and distinctions between these
two research areas will be drawn throughout this chapter.
Various explanations have been proposed for the better-than-average eff ect (see
Gilovich, Epley, & Hanko, this volume).  ese explanations encompass two broad
issues.  e rst issue concerns the role of behavioral and interpersonal comparisons
in the better-than-average eff ect . One prominent explanation for the eff ect is that
when asked to compare themselves with an average peer, people select comparison
targets who fare especially poorly on the judgment dimension (Perloff & Fetzer,
1986). Another possibility is that people think selectively about behaviors on
which they fare better than others (Weinstein, 1980). In contrast to these views,
Alicke et al. (2001) have argued that behavioral comparisons are unnecessary to
account for the eff ect, and that people routinely employ a “better-than-average”
heuristic which entails a compromise between existing self-knowledge and ideal
trait conceptions.
A second main issue underlying better-than-average explanations is whether
nonmotivational mechanisms can account for the eff ect.  e four most promi-
nent nonmotivational explanations center on whether people selectively recruit
information or comparison targets that ensure their own superiority, whether the
judgment task encourages people to focus on themselves rather than on the average
peer (focalism), whether people’s own behaviors or characteristics are considered
more thoroughly and weighted more heavily (egocentrism), and whether the
eff ect is due to diff erences between comparing a single entity (the self ) with an
aggregate (an average peer).
To anticipate our conclusion, we do not believe that nonmotivational mecha-
nisms account suffi ciently for the better-than-average eff ect, and we also believe
that various lines of evidence indicate that self-enhancement motives contribute
to the eff ect. Furthermore, many of the nonmotivational explanations that have
been discussed are as readily interpretable in motivational terms.  e fact that
people concentrate unduly on their own characteristics in making comparisons
(egocentrism), for example, could result from the tendency to believe that their
own characteristics are better or more important than others’. At the same time,
we hardly wish to argue that nonmotivational mechanisms are unimportant in
explaining the better-than-average eff ect. Any satisfactory explanation of self-
related eff ects must encompass both the why and the how of behavior. To argue
that the better-than-average eff ect occurs because people wish to view themselves
positively tells us nothing about how the eff ect occurs.  erefore, after a brief
historical survey of the emergence of better-than-average eff ect research, and a
consideration of factors that moderate the eff ect, we discuss the mechanisms that
contribute to the better-than-average eff ect. After this, we review evidence that
points to the role of self-enhancement motives in this research area and discuss
avenues for future research.
Alicke_RT19523_C05.indd 86Alicke_RT19523_C05.indd 86 7/11/2005 11:06:51 AM7/11/2005 11:06:51 AM
Data collected in conjunction with the 1976 College Board Exams provide one
of the earliest, most striking, and most frequently-cited demonstrations of the
better-than-average eff ect. Of the approximately one million students who took
the SAT that year, 70% placed themselves above the median in leadership abil-
ity, 60% above the median in athletic ability, and 85% rated themselves above
the median in their ability to get along well with others. Amazingly, 25% of the
students rated themselves in the 1st percentile on this latter characteristic.  ese
data are noteworthy because in contrast to many subsequent better-than-average
and optimistic bias studies, students were not asked to compare themselves to an
average peer but simply to indicate where they stood in relation to the median.
us, these results cannot be ascribed to negative connotations associated with
the word “average.”
Around this time, Cross (1977) distributed a questionnaire to instructors at
three branches of the University of Nebraska.  is questionnaire was concerned
primarily with undergraduate teaching issues, but included a question that asked
professors to rate their teaching abilities. Results showed that 94% of the faculty
considered themselves above average in teaching ability and 68% placed their
teaching abilities in the top 25%.  ese data demonstrated at the outset that the
better-than-average eff ect was not limited to college students.
Another frequently-cited study by Svenson (1981) showed that 88% of
American college students, and 77% of Swedish college students, considered
themselves to be above the 50th percentile on driving safety. Svenson’s research
was motivated by an earlier study in which Preston and Harris (1965) compared
50 drivers who had been hospitalized following car accidents (34 of whom had
caused the accidents, according to police records) with 50 matched drivers without
accident histories. Preston and Harris’s results showed not only that both groups
considered themselves to be above average in driving skills, but that the accident
group’s evaluation of their driving abilities did not diff er from those who were
uninvolved in accidents.
e rst experimental research on the better-than-average eff ect was con-
ducted in France where Codol (1975) studied what he called the “superior con-
formity of the self.” Codol placed his research in the context of identifying with
desirable norms.  e hypothesis guiding these studies was that people believed they
adhered to desirable norms more than others. Codol employed various self versus
other measurements in his twenty studies, and so this research did not establish a
basic paradigm for subsequent investigations. Furthermore, the context of norm
identifi cation obscured somewhat the general, social-comparative implications
of the better-than-average eff ect. Nevertheless, Codol presciently raised issues
that still resound in the better-than-average eff ect literature. His fi ndings suggest,
for example, that the better-than-average eff ect is larger when people compare
Alicke_RT19523_C05.indd 87Alicke_RT19523_C05.indd 87 7/11/2005 11:06:51 AM7/11/2005 11:06:51 AM
themselves to others in general than to specifi c group members. Codol also con-
jectured that the tendency to view oneself as superior to others represents a desire
to self-enhance rather than to denigrate others.
Self researchers recognize that people are not indiscriminately self-serving (e.g.,
Baumeister, 1998; Schlenker, 1980; Sedikides & Gregg, 2003). Self-serving
tendencies such as the better-than-average eff ect are pervasive but not inevitable.
As previously noted, even the earliest better-than-average eff ect studies assessed
moderating factors that alter the eff ect’s strength. Four main classes of moderat-
ing factors have been identifi ed: the scales on which the eff ect is measured, the
nature of the judgment dimension, the nature of the comparison target, and
characteristics of the judge.
Direct and Indirect Measurements
Research on the optimistic bias and better-than-average eff ect employs two basic
methodologies. With the direct method, self is compared to an average peer on a
single scale that uses “average” as the midpoint. At the low end, direct scales are
usually anchored with wording such as “considerably below average” and on the
high end at “considerably above average.”  e estimate of the better-than-average
eff ect is straightforward:  e higher the number circled, the greater the magni-
tude of the eff ect. With indirect ratings, participants rate the self and average
peer on separate scales.  e better-than-average eff ect is calculated by subtracting
the average rating from the self rating so that higher scores indicate greater bias.
Studies suggest that people are more self-serving when they use the direct rather
than the indirect scale (Otten & van der Pligt, 1996). Direct scales provide a
stronger comparative frame and may, therefore, elicit more pronounced tenden-
cies to contrast the self upward from the average peer or to contrast the average
peer downward from the self.
e direct method of assessing the better-than-average eff ect is used more
often, although it is less informative. With the direct method, it is impossible to
estimate whether the better-than-average eff ect results from people underestimating
the average peer’s standing, overestimating their own standing, or both.  e one
exception to this occurs when a person’s standing on a dimension is objectively
known and can be used as a reference point (Epley & Dunning, 2000).  e in-
direct method, by contrast, is informative of the direction of contrast. Because
the indirect method has been less frequently used, there is no solid basis yet for
concluding whether the better-than-average eff ect represents self-infl ation, average
peer defl ation, or some combination of both.
Alicke_RT19523_C05.indd 88Alicke_RT19523_C05.indd 88 7/11/2005 11:06:51 AM7/11/2005 11:06:51 AM
The Nature of the Judgment Dimension
People who claim positive characteristics that are easily refuted risk being ridiculed.
Furthermore, the need to maintain coherent and believable self-images (Swann,
Rentfrow, & Guinn, 2003) is threatened when people cling tenaciously to dubi-
ous abilities and characteristics. Self-serving tendencies, therefore, operate within
reality bounds. As a general rule, we assume that people are most self-serving when
they have the latitude to interpret events in a self-serving manner (Sedikides &
Strube, 1997). Self-enhancement is accomplished with the least obvious distortion
when the judgment dimension is subjective or abstract as opposed to objective or
concrete. Self-enhancement is also facilitated when people believe they have the
ability to alter their standing on the dimension. Each of these factors is discussed
separately below.
Criteria for Assessing Traits
e criteria for assessing intellectual and physical abilities are generally more
objective than those for evaluating social or moral ones (Reeder & Brewer, 1979;
Rothbart & Park, 1986).  e view that people are more self-serving when making
subjective or ambiguous judgments than objective ones leads to the prediction
that the better-than-average eff ect will be larger on ability than on social or moral
judgment dimensions.
is expectation has been confi rmed by Allison, Messick and Goethals (1989)
who found that the tendency for people to believe that they performed more
moral behaviors than their peers was greater than their tendency to believe they
performed more intellectual behaviors, although the latter was still signifi cant.
Allison et al. termed this the “Muhammed Ali eff ect.”
e ubiquitous Muhammed Ali also provides a fi tting introduction to Dun-
ning, Meyerowitz, and Holzberg’s (1989) demonstration that ambiguity moder-
ates the better-than-average eff ect. In an interview with Muhammed Ali after
winning an early fi ght, sportscaster Howard Cosell suggested that Ali was mighty
“truculent” that evening, to which Ali replied: “I don’t know what truculent is,
but if it’s good, I’m it.” Dunning et al. captured something like this reasoning
in a more formal and less truculent manner.  eir rst two studies showed that
the better-than-average eff ect was greater on dimensions that had been preclas-
sifi ed as ambiguous versus unambiguous. In the third study, trait ambiguity was
manipulated by presenting some participants with specifi c criteria for assessing a
trait, whereas others were free to defi ne the traits for themselves. Results generally
showed that the better-than-average eff ect was larger when participants provided
their own trait defi nitions, although this eff ect was more consistent for positive
than for negative trait dimensions.
Alicke_RT19523_C05.indd 89Alicke_RT19523_C05.indd 89 7/11/2005 11:06:51 AM7/11/2005 11:06:51 AM
In addition to having the latitude to interpret a trait’s meaning, self-enhancement
is facilitated when people can construe their standing on a trait in a self-serving
manner. Positive characteristics that people believe they control have greater self-
serving value than characteristics they believe are less alterable, whereas negative
uncontrollable characteristics are less defl ating than controllable ones.
e rst large-scale, systematic study of the better-than-average eff ect assessed
the moderating infl uence of a trait characteristic’s perceived controllability. Partici-
pants in this study examined self versus average ratings on separate scales for 171
trait dimensions (Alicke, 1985).  ese traits were prerated to represent four levels
of desirability (high, moderately-high, moderately low, and low) and two levels
of controllability (high and low).  e larger set of desirability than controllability
categories refl ects the greater range in desirability preratings.
Although we assumed that people would evaluate themselves more favorably
than they would an average peer, we expected this tendency to be moderated by
controllability.  e primary prediction was that participants would believe them-
selves to be characterized more by positive controllable than positive uncontrollable
traits in relation to the average college student, and more by negative uncontrol-
lable than negative controllable traits.  ese predictions can be summarized in
the phrase: “I make me good, fate makes me bad.”
As we anticipated, the tendency to evaluate oneself more favorably than
the average college student on positive traits, and less unfavorably on negative
traits, was pervasive.  e predicted eff ects of controllability were also obtained
such that participants rated themselves more favorably in relation to the average
college student on positive controllable traits and more unfavorably on negative
uncontrollable traits.
The Nature of the Comparison Target
A fundamental question surrounding better-than-average eff ect judgments con-
cerns the nature of the comparison target. Whereas traditional social comparison
studies include comparisons between individuals, better-than-average eff ect
research entails comparisons between oneself and an hypothetical or statistical
entity, namely, an average peer. Extensive attributional and decision-making
research shows that people tend to deemphasize or misuse statistical information
(Nisbett & Ross, 1980).  e better-than-average eff ect, therefore, might disap-
pear when comparisons are eff ected between real people rather than between a
person and a statistic.
Alicke et al. (1985) conducted a series of studies to see if the better-than-
average eff ect would be eliminated when people compared themselves to a real
person rather than an average peer. In their fi rst and simplest study, half the
participants were brought to a large room and asked to look at the person sit-
ting next to them.  ese participants then changed their seats and made 40 trait
Alicke_RT19523_C05.indd 90Alicke_RT19523_C05.indd 90 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
comparisons (20 positive, 20 negative) between themselves and the person they
had sat next to. A second group of participants compared themselves on these
same dimensions to the average college student. Results showed that the better-
than-average eff ect was pervasive in both groups, but was signifi cantly reduced in
comparisons with real people.
is rst study suggested that people adjust their evaluations when comparing
with real versus hypothetical targets, but still view themselves more favorably in
person-to-person comparisons. Six more studies were conducted to investigate in
greater detail the diff erences between real and hypothetical comparison targets.
ese studies used a common paradigm in which an interviewer asked a series of
predetermined questions of an interviewee (actually a confederate) who always
gave the same stock answers. In the fi rst study using this paradigm, a live observer
watched the interaction in the same room, another group watched the interaction
on videotape, a third read a written transcript of the interview, and a fourth made
self versus average college student judgments. Ratings in the fi rst study were made
on the kinds of life events studied in optimistic bias research. Consistent with the
rst study’s fi ndings, the better-than-average eff ect (or optimistic bias in this case)
was greater when participants compared themselves to the average college student
than in any of the other conditions.  e more novel fi nding of this study was that
the better-than-average eff ect was greater in the transcript and video conditions
than in the live observer or interviewer conditions. No diff erences were obtained
between the live observer and interviewer conditions, suggesting that actual
interaction with the target does not infl uence comparisons beyond experiencing
the target’s live presence.
is study, therefore, established two diff erences between real and hypotheti-
cal comparison targets.  e rst diff erence is individuation. Any specifi c target
ostensibly reduces the better-than-average eff ect in relation to comparisons with an
hypothetical entity such as an average peer.  e second diff erence is live contact.
e better-than-average eff ect is reduced when people are in the same room with
the comparison target regardless of whether an actual interaction takes place.
Subsequent studies sought further refi nements. Participants in the individu-
ation condition of the previous study received some information from the target
in the form of the target’s answers to the interview questions. To create even more
basic individuation conditions, participants in one group saw only a still image
of the target, and in another, saw only the back of the target’s head (to eliminate
facial cues). We also created conditions in which participants thought they were
watching a contemporaneous interview on a TV monitor, and conditions in
which participants watched the interview from behind a one-way mirror with
the belief that the interviewee could, or could not, see them. In the mirror condi-
tions, participants stood almost the exact distance from the interviewee as in the
live observer condition, and also saw the interviewee from the same angle.  e
results were clear: Every condition in which participants compared themselves
to the interviewee produced a decreased better-than-average eff ect in relation to
Alicke_RT19523_C05.indd 91Alicke_RT19523_C05.indd 91 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
comparisons with an average college student, and an increased better-than-average
eff ect in relation to live observer conditions.
ese ndings suggest that individuation per se reduces the better than aver-
age eff ect.  e results of these studies also show that the features that diff erentiate
live- from nonlive contact are quite subtle. Live contact did not increase feelings
of similarity to the interviewee, nor did it require any type of interaction. Simply
being in the same room with the target was suffi cient to reduce the better-than-
average eff ect, and this same reduction did not occur when participants believed
they were watching the interaction live on a monitor, or even when they watched
the interview through a mirror in the next room and knew that the interviewee
could see them.
Another issue these studies investigated was whether participants perceived
the “average” student pejoratively. One possible explanation for better-than-average
eff ect fi ndings is that people do not want to be considered average because of its
negative connotation. To assess how participants viewed the average student, we
had them create distributions for 16 diff erent trait dimensions. For example, for
the trait dimension dependable-undependable, participants listed the percentage
of people they thought fell into nine categories between extremely dependable
and extremely undependable, with the understanding that their percentages
should total to 100%.  e mean of each trait dimension was calculated, and this
value was compared to where on the dimension participants placed themselves,
the average college student, or a real person whom they had sat next to. As in
the previous studies, participants evaluated themselves and the real person more
favorably than the average college student, while consistently placing themselves
above the real person. More germane for the purposes of this study, participants
consistently placed the average college student above the distribution mean. In
these data, therefore, the average college student was not viewed pejoratively, at
least not in relation to the distribution mean.  ese ndings suggest that even the
average college student is perceived as a more individuated entity than the mean
of a trait distribution.
Characteristics of the Judge
Relatively few better-than-average eff ect studies have examined individual dif-
ference factors.  e one factor that has been routinely analyzed–gender–rarely
produces signifi cant eff ects. In this section, we briefl y review the two characteristics
that have received some attention, namely self-esteem and depression.
Not everyone believes they exceed the average by the same degree. Self-esteem is
perhaps the fi rst individual diff erence factor that comes to mind in considering
variations in the better-than-average eff ect. In fact, self-esteem did come to mind
very early in research on this topic. Brown (1986) found that the tendency to
Alicke_RT19523_C05.indd 92Alicke_RT19523_C05.indd 92 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
evaluate oneself more favorably than others was greater for high self-esteem than
for low self-esteem participants, although this eff ect was obtained for positive and
not for negative traits.
More recently, Suls, Lemos, and Stewart (2002, Study 1) assessed the self
versus average peer comparisons of high and low self-esteem participants on traits
varying in ambiguity.  ey found that whereas both high and low self-esteem
individuals exhibited a greater better-than-average eff ect for ambiguous versus
unambiguous traits on positive trait dimensions, low self-esteem individuals did
not show this ambiguity eff ect on negative trait dimensions.  us, only high
self-esteem individuals took advantage of the interpretational latitude aff orded
by negative, ambiguous traits.
Tabachnik, Alloy, and Crocker (1983) compared the self versus average peer judg-
ments made by students who scored relatively high or low on the Beck Depression
Inventory.  eir main hypothesis was that those who scored higher would view
themselves as more similar to the average college student on depression-relevant
items but not on irrelevant items. As it turned out, depressive participants viewed
themselves as more similar to average on both depression-relevant and depression-
irrelevant items. Because the depression-relevant items were all negative, and the
depression-irrelevant items were predominantly positive, these fi ndings suggest that
depressives exhibit a diminished better-than-average eff ect across the board. In other
words, depressives have a reduced tendency to evaluate themselves less negatively on
negative characteristics relative to the average student as well as a reduced tendency
to evaluate themselves more positively on positive characteristics.
Five primary mechanisms have been proposed to explain how the better-than-aver-
age eff ect operates. One prevalent idea is that people selectively recruit downward
targets who make them look favorable by comparison, or relatedly, that they
selectively recruit behavioral evidence that favors the self. A second prominent
explanation is that people focus egocentrically on their own positive attributes and
that the heightened availability of their own behaviors and propensities produces
the better-than-average eff ect.  ird, focusing explanations argue that the position
of the self as the subject of judgment and the average person as the target produces
the better-than-average eff ect. By this reasoning, reversing the position of subject
and target should eradicate the eff ect. Fourth, the self versus aggregate position
argues that individual entities, such as the self, are evaluated more favorably than
group or aggregate estimates, such as an average peer. Finally, the better-than-
average eff ect could be a heuristic that is applied automatically in social judgments
and then modifi ed for specifi c comparison targets or dimensions. Each of these
explanations is discussed in turn below.
Alicke_RT19523_C05.indd 93Alicke_RT19523_C05.indd 93 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
Selective Recruitment
Most explanations of the optimistic bias, and some of the better-than-average
eff ect, involve the way people think about their characteristics in relation to
others. In his early optimistic bias research, Weinstein (1980, 1984; Weinstein &
Lachendro, 1982) proposed the most prevalent variant of this explanation, namely,
that when people compare their characteristics to others, they think selectively
about their own strengths or about others’ weaknesses. Weinstein fi rst tested the
selective recruitment hypothesis in a study (1980, Study 2) in which participants
listed behaviors that increased or decreased their chances of experiencing each of
a series of life events. Some participants were then given the opportunity to read
others’ lists. Results showed a reduced optimistic bias in participants who read
others’ lists versus those who did not have access to this information. Importantly,
access to other people’s responses reduced, but did not eliminate, the optimistic
bias. Weinstein and his colleagues showed similar reductions in the optimistic bias
in studies that provided participants with specifi c information about others’ risks
for misfortune (Weinstein, 1984; Weinstein & Lachendro, 1982).
Perfl off and Fetzer (1986) considered another aspect of the selective recruit-
ment hypothesis, namely, that when asked to compare themselves with an average
peer, people select targets who compare unfavorably on the judgment dimension.
People may think, for example, of an especially dishonest person, which casts
their own honest behaviors in an especially favorable light. To test this downward
comparison idea, Perloff and Fetzer had participants compare their vulnerabilities
to misfortune with those of their closest friend, a close friend, and the average
college student. Perloff and Fetzer assumed that identifying a specifi c, well-known
comparison target (i.e., their closest friend) would prevent participants from se-
lecting a target who was worse off than themselves on the comparison dimension
or from recruiting specifi c behaviors or characteristics on which they fare better.
Consistent with this assumption, they found that what they called “the illusion
of invulnerability” was reduced when people compared themselves to their clos-
est friend, relative to when they compared with a close friend or with an average
college student.
As Perloff and Fetzer noted, however, there are competing explanations for
these fi ndings.  e explanation they favored was that people possess more infor-
mation about their closest friends, which enables them to conclude that these
friends are no more susceptible to misfortune than themselves. Another plausible
explanation, however, is that people like their closest friend more than a close
friend or an average peer and evaluate their closest friend more favorably on this
basis.  ese studies, therefore, provide less clear evidence about the moderating
role of behavioral information than Weinstein and his colleagues’ research (1980
1984; Weinstein & Lachendro, 1982). What Perloff and Fetzer’s results do suggest
is that the better-than-average eff ect is reduced when positive self-evaluations are
extended to others, such as close friends.
Alicke_RT19523_C05.indd 94Alicke_RT19523_C05.indd 94 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
Egocentrism is the probably the most prevalent nonmotivational explanation
for better-than-average and optimistic bias eff ects. Egocentrism as applied to the
better-than-average eff ect is the tendency to place undue weight on one’s own
characteristics, beliefs and experiences in making self versus average comparisons.
In contrast to the selective recruitment hypothesis, egocentrism does not necessarily
entail a self-serving review of behavioral evidence. In judging their relative honesty,
for example, people may consider the same honest behaviors for themselves and
the target but still place greater weight on their own honest behavior. Furthermore,
selective recruitment can entail thinking about the other’s negative characteristics
without focusing unduly on one’s own.
One source of support for the egocentrism view comes from studies show-
ing that self versus average peer comparisons are predicted better by absolute self
ratings (that is, self ratings alone, without ratings of the average) than by absolute
peer ratings (that is, peer ratings alone, without ratings of the self). Klar and
Giladi (1999), for example, had participants make absolute ratings of their own
contentment, absolute ratings of their peers’ contentment, and also comparative
ratings of their own contentment relative to their peers.  e main nding in their
two studies was that absolute self-ratings predicted the comparative contentment
ratings better than did absolute peer ratings. In fact, the relationship between
absolute peer ratings and the comparative ratings were low and nonsignifi cant in
both studies. Although these studies examined only one trait dimension, other
studies have obtained analogous results with diff erent judgment tasks (e.g., Eiser,
Pahl, & Prins, 2001; Chambers, Windshitl, & Suls, 2003).
One of the most compelling demonstrations of the egocentrism position
is Kruger’s (1999) fi nding that people consider themselves worse than average
on diffi cult tasks. Kruger reasoned that if concentrating egocentrically on their
positive attributes leads people to think that they are better than average, then
concentrating on their negative attributes should lead them to believe that they
are worse than average.  is prediction can also be viewed from an anchoring and
adjustment perspective: In the case of tasks for which people believe that they
have high ability, anchoring on their own characteristics should lead to relatively
extreme positive self-judgments, with insuffi cient upward adjustments for their
peers, whereas for tasks on which people believe that they have low ability, anchor-
ing should produce extreme negative self-judgments, with insuffi cient downward
adjustments for their peers.
Based on pretesting, Kruger classifi ed activities as easy (e.g., driving, using
a mouse) or diffi cult (telling jokes, juggling) and then had participants estimate
their percentile ranking for each of the activities. In accord with the egocentrism
position, participants consistently placed themselves above the 50th percentile for
easy activities, and below the 50th percentile for diffi cult ones. ese ndings are
consistent, therefore, with the assumption that people concentrate egocentrically
Alicke_RT19523_C05.indd 95Alicke_RT19523_C05.indd 95 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
on their own attributes in comparative judgments and that emphasis on their
negative characteristics leads them to overestimate their shortcomings.
e tendency to concentrate egocentrically on personal prospects and charac-
teristics has important implications for self-other comparisons. In general, if people
think egocentrically about their own prospects, then factors that increase their
chances of success at the task should induce overconfi dence about their prospects
(because people focus egocentrically on their own advantage without realizing
others have the same advantage), whereas factors that augur equally unfavorably
for themselves and others should lead to pessimistic predictions. Chambers,
Windschitl, and Suls (2003) tested this hypothesis by asking participants to predict
the likelihood that they versus an average peer would purchase their dream home
within a short time frame (next 6 years) or a long one (next 32 years). Because the
probability is higher that the event will occur in the longer time frame, egocentrism
predicts that people will be overly optimistic about their chances in the long than
in the short time frame.  e results confi rmed this prediction.
Focalism is the tendency to place greater weight on whatever hypothesis or outcome
is currently the focus of attention (Schkade & Kahneman, 1998). In contrast to
egocentrism, which explicitly involves self-reference, focalism involves concentrat-
ing on an object due to the way a judgment task is structured. By asking people
to compare their characteristics to those of an average peer, studies on the better-
than-average eff ect tend to place the self in the focal position and the average peer
in the referent position. Because self-representations contain a greater number
of unique qualities than other representations (Karylowski, 1990; Karylowski &
Skarzynaka, 1992), focusing on the self highlights these unique features and leads
people to perceive themselves as less similar to the average.
By making the self the focal object, therefore, the better-than-average eff ect
methodology increases the perceived diff erences between self and other. According
to this reasoning, when people compare the average other to themselves, these
diff erences should be attenuated. In other words, if the positions of self and aver-
age are switched, such that the average peer is made the focal object and the self
is made the referent, the better-than-average eff ect should be reversed or at least
e main support for this focalism prediction comes from studies using the
optimistic bias paradigm. Otten and van der Pligt (1996) and Eiser, Pahl, and
Prins (2001) both manipulated whether participants were asked to estimate how
they would fare relative to their peers on various life events (self-other focus), or
how their peers would fare relative to themselves (other-self focus).  ese studies
showed a reduced optimistic bias in the latter condition, that is, when the average
peer was the focal object and the self was the referent.
Alicke_RT19523_C05.indd 96Alicke_RT19523_C05.indd 96 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
Other studies have compared focalism and egocentrism predictions, although
not with average peer comparisons. In two similar lines of research, Windschitl,
Kruger, and Simms (2003) and Moore and Kim (2003) placed participants in a
competitive situation and asked them to estimate their chances of success.  ese
studies assessed egocentrism by varying whether participants believed the task
facilitated or discouraged success (e.g., playing an easy or hard trivia game). Be-
cause the task was equally diffi cult for themselves and their opponent, there was
no rational reason for them to alter their estimates based on this information. But
from the egocentrism standpoint, concentrating disproportionately on one’s own
prospects should lead to overestimation in the case of an easy task and underes-
timation in the case of a diffi cult one, which is what these studies demonstrated.
Focalism was independently manipulated by asking participants to estimate their
own or their opponent’s chances of winning. Although the results varied some-
what across experiments, both focalism and egocentrism infl uenced participants’
estimates of success.
Self Versus Aggregate Comparisons
In the better-than-average and optimistic bias paradigms, a single entity, the self,
is compared to an aggregate, the average peer.  e fact that the self is routinely
evaluated more favorably than average is generally believed to manifest self-es-
teem enhancement. However, Klar and Giladi’s demonstration of “non-selective
superiority and inferiority” biases (Klar, 2002; Klar & Giladi, 2002) calls into
question whether self-enhancement assumptions are needed to explain the better-
than- average eff ect. What Klar and Giladi have demonstrated in numerous experi-
ments is that any member of a positively-evaluated group is rated more favorably
than the group average (Klar, 2000; Klar & Giladi, 1997). Randomly-selected
students at one’s university, for example, are evaluated more favorably than the
average student at the university.  is nding obtains even when comparing an
individual group member to other distinct individuals, such as comparing a single
police offi cer to the average of other police offi cers in the room. Giladi and Klar
(2002) have demonstrated this same eff ect with impersonal comparisons, such as
soap fragrances and musical selections.
Klar and Giladi’s (1997, 2002) fi ndings suggest that the greater positivity
people claim for themselves may be subsumed by a more general tendency to place
greater weight on single entities than on aggregates.  e generality of their view
is extended by their consistent fi ndings of inferiority biases, that is, the tendency
for members of disliked groups to be evaluated less favorably than the group as
a whole.
Klar and Giladi’s fi ndings are consistent with those of Alicke et al. (1995) in
showing that the better-than-average eff ect is reduced by comparisons with indi-
viduated entities versus an average peer. Klar and Giladi’s results suggest further
Alicke_RT19523_C05.indd 97Alicke_RT19523_C05.indd 97 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
that part of the tendency to evaluate oneself more favorably than an average peer
is due to the greater weight people place on any individuated entity versus an ag-
gregate such as an average peer. On the other hand, Alicke et al.’s research show
that compared to other individuated entities, the self has a privileged role in that
the better-than-average eff ect is greater when the self is compared to any other
individuated entity.  us, while Klar and Giladi’s model provides a cogent and
general account of individual-group comparisons, an additional factor appears to
be operating when the self is plugged into the comparison.
Klar and Giladi have recently expanded their view in what they call the
LOGE model (local comparisons-general standards model, 2002). According
to this model, the task of estimating, for example, a group member’s politeness
relative to the average student, requires comparing the individual’s politeness to a
local standard, namely, the average level of politeness in the immediate peer group
(for example, students at this university). When people make this comparison,
however, they are unable to avoid applying a more general standard, which might
include all other people. To the extent that the local standard is more favorable
than the general one (i.e., this group is more polite than people in general), supe-
riority biases should emerge such that any person in the group will be evaluated
more favorably than the group average.  is occurs because people inadvertently
take into account the superiority of the local standard to the general one, rather
than simply recognizing that the person is an average member of a superior group.
By this same reasoning, evaluations of any individual who belongs to an inferior
group (relative to the general standard) should be less favorable than the group
average.  e LOGE model, therefore, provides a useful and general account of
comparisons between specifi c entities and group averages.  e model’s limitation
as applied to the better-than-average eff ect is that it does not contain mechanisms
to explain the enhanced favorableness that is generally accorded to the self versus
other entities.
Better-Than-Average Heuristic
Research on selective recruitment leaves little doubt that the optimistic bias is
altered by providing people with access to others’ beliefs about their prospects
in life. We question, however, whether careful thinking about one’s behavior is
a necessary, or even a typical component, of self versus other comparisons.  e
assumption that people think carefully about specifi c behaviors is less tenable in
the better-than-average eff ect paradigm than in optimistic bias research. In better-
than-average eff ect research, participants typically judge abstract traits rather than
concrete behaviors. Furthermore, the better-than-average eff ect has been obtained
in settings in which participants make hundreds of trait comparisons, and it seems
unlikely that they engage in careful behavior analyses for each comparison.
Alicke et al. (1995; 2001) have suggested that the better-than-average eff ect
is attributable to people applying a better-than-average heuristic.  is heuristic
Alicke_RT19523_C05.indd 98Alicke_RT19523_C05.indd 98 7/11/2005 11:06:52 AM7/11/2005 11:06:52 AM
entails an automatic tendency to assimilate positively-evaluated social objects
toward ideal trait conceptions, and does not assume that people routinely review
their behaviors to make self-other judgments.  e assumption that people apply
a better-than-average heuristic is consistent with Sears’ (1983) notion of a person
positivity bias, and with the general positivity bias that pervades social judgment
(Matlin & Stang, 1978).  e degree of assimilation varies for social objects of dif-
ferent value. Family members and friends are accorded a great deal of positivity,
and concrete individuals are accorded more than an average or hypothetical peer.
At the apex of the positivity ladder resides the self.
e extent to which people assimilate toward ideal trait conceptions depends
on the ambiguity of the judgment dimension and on the strength of prior self-
conceptions. As noted previously, people are not indiscriminately self-serving and
tend to avoid easily-refutable claims. Nevertheless, trait comparisons are especially
susceptible to the better-than-average heuristic because trait conceptions can be-
come independent of behavioral exemplars (Klein & Loftus, 1993; Klein, Loftus,
& Burton, 1989; Klein, Loftus, Trafton, & Furman, 1992). Research by Klein and
Loftus shows that people require the same amount of time to recall an instance in
which they displayed a trait regardless of whether they fi rst judge whether that trait
is self-descriptive or simply defi ne the trait. If people accessed specifi c behaviors
to answer trait questions, then judging whether a trait was self-descriptive would
facilitate recalling an instance in which the trait was displayed. Based on numer-
ous failures to fi nd such facilitation eff ects, Klein and Loftus argue that trait and
behavioral information are stored in separate memory systems.
e better-than-average heuristic entails three main assumptions. First, when
people are asked, for example, to judge their “kindness” in relation to an average
peer, the default is to assimilate their self-ratings toward their ideal conceptions of
kindness.  ese ideal trait constructs do not necessarily translate into the highest
available scale point. People who are too cooperative, for example, can be taken
for patsies, and extreme honesty can elide into rudeness.
e second assumption is that people make automatic adjustments based on
past self-conceptions.  ose who have frequently been criticized for their unhelp-
fulness will still associate with ideal conceptions of helpfulness but will assimilate
less to accommodate reality.  e nal assumption is that average peers, rather
than being assimilated toward idea standards, are evaluated in relation to oneself.
Because the self typically represents a relatively high scale point, average peers are
assimilated toward oneself, while still being rated less favorably.
Although better-than-average heuristic assumptions have not been tested
directly, there is strong evidence to suggest that behavior recruitment is not a
necessary component of the better-than-average eff ect. One source of support for
this assertion comes from the fact that the better-than-average eff ect emerges even
under extreme cognitive load conditions (Alicke et al., 1995, Study 7).
Another source of support for the nonbehavioral assumption comes from re-
search on what we have called the “better-than-myself” eff ect (Alicke et al., 2001).
Alicke_RT19523_C05.indd 99Alicke_RT19523_C05.indd 99 7/11/2005 11:06:53 AM7/11/2005 11:06:53 AM
In our fi rst study on this topic, participants in a pretesting session estimated the
percentage of times they exhibited behaviors relevant to various trait dimensions.
For example, participants were asked to estimate the percentage of times they
were cooperative or uncooperative when the opportunities to display that trait
arose.  is behavior percentage methodology was modeled on the act-frequency
approach to personality (Buss & Craik, 1983) which assumes that people defi ne
their traits by estimating the frequency with which they engage in trait-relevant
behaviors. Participants in Study 1 were told to use their percentage estimates to
rate themselves on each corresponding trait dimension.
In the main session conducted approximately six weeks later, participants
received what they believed were the average behavior percentage estimates ob-
tained during the academic quarter. What participants actually received were the
identical estimates they had provided in the pretesting session.  us, if participants
estimated that they were cooperative 86% of the time and uncooperative 14% of
the time, they were led to believe that the average student was cooperative 86%
of the time and uncooperative 14% of the time. Participants were asked to use
these estimates to evaluate where they and the average college student fell on the
trait dimension.
Results were consistent across the board: Despite looking at the exact behav-
ior estimates they had provided in pretesting, participants evaluated themselves
more favorably than the average college student on almost every dimension.
ese ndings were replicated in a second study in which participants received
what they believed were the behavior estimates made by a randomly-selected peer
rather than the average college student. Although the magnitude of the eff ect was
reduced, participants still placed themselves signifi cantly above their peers based
on identical behavior estimates.
A third study assessed whether participants might want to change their be-
havior estimates once they saw the estimates of an average person or a peer. A pos-
sible explanation for the previous studies’ results is that participants believed they
had underestimated the frequency with which they engaged in positive behaviors
after seeing others’ estimates. To test this, we gave participants the opportunity
to change their frequencies after seeing others’ estimates. In general, participants
made relatively few changes. Furthermore, changes that were made did not cor-
relate with comparative ratings.
e better-than-myself paradigm used in these studies has one notable
limitation, namely, that while participants might readily acknowledge that their
behavior frequencies are similar to others’, they could still conclude that their own
trait-relevant behaviors are more exemplary. For example, people might accept that
they and another person are cooperative 85% of the time but believe that their
own cooperative behaviors are more cooperative than someone else’s. We used a
diff erent methodology to circumvent this problem in a fourth study.  is time, we
asked participants to list every behavior they could think of that refl ected where
they stood on one of four trait dimensions (kind-unkind, intelligent-unintelligent,
Alicke_RT19523_C05.indd 100Alicke_RT19523_C05.indd 100 7/11/2005 11:06:53 AM7/11/2005 11:06:53 AM
honest-dishonest, and creative-uncreative ). After listing the relevant behaviors,
participants received a list made by another student and then compared themselves
to the student on the trait dimension. On average, the lists participants received
from others should have been just as positive as the ones they produced themselves.
Nevertheless, with a peer’s trait-relevant behaviors in front of them, they continued
to evaluate themselves more favorably than the peer. We believe that this study
provides reasonably strong evidence that diff erences in behavior recruitment are
not a necessary component of the better-than-average eff ect.
Selective recruitment, focalism, and egocentrism have all been shown to moder-
ate the better-than-average eff ect. ese judgment features have been proposed
as alternatives to self-enhancement assumptions (Chambers & Windshitl, 2004).
e credibility of the nonmotivational position would be heightened if it could
be shown that these factors, either in isolation or combination, eliminate the
better-than-average eff ect. But as a general rule, variations in these judgment facets
alter, but do not eliminate, the better-than-average eff ect. For example, for focal-
ism to provide a suffi cient explanation of the better-than-average eff ect, people
must evaluate average peers more favorably than themselves when the average
peer is the focal object and the self is this referent.  is is not what happens. In
studies on focalism, reversing the position of self and average attenuates but does
not eliminate the eff ect.  e same is true for eff ects attributable to egocentrism
or selective recruitment.  us, the specifi c information people focus on, and the
kinds of comparisons they make, while important moderators of the better-than-
average eff ect, do not suffi ce to explain it.
is failure of these various mechanisms to account completely for the better-
than-average eff ect does not, of course, establish the role of self-enhancement. But
various other fi ndings do suggest a role for self-enhancement.  e nding in our
early study (Alicke, 1985) that the better-than-average eff ect increases with posi-
tive controllable traits and decreases with negative uncontrollable traits, provides
one source of support for the self-enhancement motive.  is result shows that
people are most self-aggrandizing when they feel responsible for their positive
characteristics, and least self-aggrandizing when they believe that fate accounts for
their negative characteristics.
at the tendency to evaluate oneself more favorably than others increases
with the desirability of the judgment dimensions provides even more basic support
for the self-enhancement motive (Weinstein, 1980; Hayes & Dunning, 1996) .
Another aspect of the better-than-average eff ect that is diffi cult to account for
without reference to self-enhancement is the consistent fi nding that the eff ect is
stronger on ambiguous or subjectively-defi ned dimensions (Allison, Messick, &
Goethals, 1989; Dunning, Meyerowitz, & Holzberg, 1989). Apparently, people
Alicke_RT19523_C05.indd 101Alicke_RT19523_C05.indd 101 7/11/2005 11:06:53 AM7/11/2005 11:06:53 AM
are most self-serving when they have the latitude to construe comparisons in a
manner that emphasizes their superiority.
Egocentrism and selective recruitment, the two most prominent and general
nonmotivational explanations of the better-than-average eff ect, assume that the ef-
fect involves the type of behaviors or comparison targets people think about, or the
relative emphasis they place on their own actions and characteristics. Research on
the better-than-myself eff ect, however, shows that the tendency to evaluate oneself
more favorably than others perseveres even when behavioral evidence is equated
for self and other. Furthermore, the tendencies to emphasize one’s own actions and
characteristics, and to recruit selectively information that casts oneself in the most
favorable light, are readily interpretable as serving the need to self-enhance.
e idea that people automatically identify with ideal trait conceptions is
seemingly contradicted by Kruger’s fi ndings of a worse-than-average eff ect and
Klar and Giladi’s fi ndings of inferiority biases.  is apparent discrepancy can be
readily resolved, however, by expanding the better-than-average heuristic view to
include the possibility for contrast as well as assimilation eff ects. Contrast eff ects
are likely to occur when the object of judgment is obviously unfavorable, such
as a behavioral weakness or a disliked individual. Kruger, for example, obtained
his worse-than-average eff ects with behaviors such as juggling and playing chess–
behaviors for which the majority of people readily recognize their shortcomings.
Instead of automatic assimilation to ideal trait conceptions, we assume that people
automatically contrast themselves from the ideal under such circumstances.
People like to think favorably of themselves, and for good reason. Positive self-views
promote harmonious personal relationships and successful goal-striving.  ose who
feel good about themselves are less prone to negative moods and depression (Taylor
et al., 2003).  e ways in which people strive to maintain favorable self-images are
legion, including taking credit for positive outcomes and denying responsibility
for negative ones (Bradley, 1978; Zuckerman, 1979), selectively recalling favorable
information about themselves (Sedikides & Gregg, 2003), exaggerating the ability
of people who outperform them and who they outperform (Alicke et al., 1997),
searching selectively for information that confi rms a positive self-image, evaluat-
ing others in a way that refl ects favorably on one’s own performance (Dunning
& Cohen, 1992), and affi rming threatened aspects of self (Steele, 1978). Each of
these behavior tendencies, either strategically or inadvertently, serves to promote
favorable self-views.
e better-than-average eff ect is diffi cult to locate in this “zoo” (Tesser,
2000) of self-enhancement mechanisms. For one thing, it is unclear whether
the better-than-average eff ect refl ects an already favorable self-image, or is con-
structed spontaneously. In other words, the better-than-average eff ect could be
Alicke_RT19523_C05.indd 102Alicke_RT19523_C05.indd 102 7/11/2005 11:06:53 AM7/11/2005 11:06:53 AM
a consequence of the aforementioned self-enhancement mechanisms, or it could
be a distinct mechanism in it own right. Although numerous better-than- average
eff ect studies have been conducted, we still do not know precisely what kind of
eff ect it is. Does the better-than-average eff ect, for example, primarily refl ect a
tendency to contrast oneself upward from the average, to contrast the average
downward from the self, or as the better-than-average heuristic implies, upward
assimilation of both self and average toward an ideal trait concept, with greater
assimilation for the self. To answer this basic question requires a design in which
diff erent groups of participants make either absolute self-judgments or absolute
average-peer judgments, followed by comparative self versus average judgments.
is design would make it possible to analyze precisely assimilation and/or contrast
eff ects in self versus average peer ratings.
e better-than-average eff ect would be less important if it were due solely
to the vague and amorphous nature of comparisons with an “average peer.” But
numerous studies have shown that people also evaluate themselves more favor-
ably than specifi c peers, although the eff ect is attenuated in such comparisons. An
interesting off shoot of better-than-average eff ect research concerns the nature of
the diff erence between comparisons with specifi c and average peers. One possible
diff erence is that people confer “personhood” on real human beings and evaluate
them more favorably than statistical entities on this basis (Sears, 1983). A related
possibility is that people are more modest in comparisons with real individuals
and therefore inhibit self-serving tendencies.
One important direction for future research is, as noted above, to compare
conditions in which people make comparative self versus average peer ratings to
those in which they rate self and average individually.  is design would answer a
fundamental question regarding the better-than-average eff ect, namely, whether
self or average ratings are altered when made comparatively, and if so, in which
direction this alteration occurs. Most researchers, including us, assume that the
self is an anchor point against which average peer ratings are referred, but this
assumption has not been tested directly in previous research. If the self-anchoring
assumption is correct, then self-ratings should not change when they are made indi-
vidually versus when they are made in comparison to the average peer. A downward
contrast of the average peer from the self would indicate that the eff ect involves
downplaying others’ characteristics relative to one’s own. A diff erent possibility
is that people anchor on the average peer, and contrast the self upward from that
point, suggesting self-infl ation relative to the average standard. A third possibility,
one that the better-than-average heuristic predicts, is that self ratings represent a
stable (and high) anchor based on immediate associations with an ideal standard,
and that comparisons with the high self standard lead to upward assimilation of
the average peer, but an assimilation that falls short of the self.
Another direction for future research is to introduce manipulations designed
to alter the better-than-average eff ect. An obvious possibility is to introduce threats
to one’s perceived standing on a trait dimension. Self-enhancement perspectives
Alicke_RT19523_C05.indd 103Alicke_RT19523_C05.indd 103 7/11/2005 11:06:53 AM7/11/2005 11:06:53 AM
predict that the better-than-average eff ect should be increased when people con-
front a threat to an important aspect of their identities. Again, this fundamental
assumption has yet to be tested explicitly.
As noted at the outset, the better-than-average eff ect is a type of social
comparison in which people are asked to evaluate themselves with reference to a
normative standard, namely, an average peer or the midpoint of a distribution.
Research on this topic has shown consistently that people place themselves above
this standard, and also above specifi c peers.  e better-than-average eff ect tells us
that people evaluate themselves more favorably than others, and this eff ect is not
due solely to the weight they place on their own characteristics in comparative
judgments, their tendencies to focus on themselves as the judgment object, or
on the tendency to recruit favorable information about themselves. In fact, one
can reasonably argue that both egocentrism and selective recruitment serve self-
enhancement needs. In other words, thinking egocentrically about one’s own posi-
tive qualities, or selecting downward comparison targets, may represent motivated
propensities to reach favorable conclusions about one’s standing relative to others.
us, various fi ndings suggest that the better-than-average eff ect is due, at least
in part, to a desire to view oneself in a favorable light relative to one’s peers.  e
task in future investigations is to evaluate which kinds of self-threats infl uence self
versus average judgments, and to assess whether such alterations entail changes in
self ratings, average ratings, or both.
Alicke, M. D. (1985). Global self-evaluation as determined by the desirability and controllability of trait
adjectives. Journal of Personality and Social Psychology, 49, 1621–1630.
Alicke, M. D., Klotz, M. L., Breitenbecher, D. L., Yurak, T. J., & Vredenburg, D. S. (1995). Personal
contact, individuation, and the better-than-average eff ect. Journal of Personality and Social Psychology,
68, 804–825.
Alicke, M. D., LoSchiavo, F. M., Zerbst, J., & Zhang, S. (1997).  e person who outperforms me is a
genius: Maintaining perceived competence in upward social comparison. Journal of Personality and Social
Psychology, 72, 781–789.
Alicke, M. D., Vredenburg, D. S., Hiatt, M., & Govorun, O. (2001).  e “better than myself ” eff ect.
Motivation and Emotion, 25, 7–22.
Allison, S. T., Messick, D. M., & Goethals, G. R. (1989). On being better but not smarter than others:
e Muhammad Ali eff ect. Social Cognition, 7, 275–296.
Baumeister, R. F. (1998).  e self. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), e handbook of social
psychology (4th ed., pp. 680–740). Boston: McGraw-Hill.
Bradley, G. W. (1978). Self-serving bias in the attribution process: A reexamination of the fact or
ction question. Journal of Personality and Social Psychology, 36, 56–71.
Brown, J. D. (1986). Evaluations of self and others: Self-enhancement biases in social judgments. Social
Cognition, 4, 353–376.
Buss, D. M., & Craik, K. H. (1983).  e act frequency approach to personality. Psychological Review, 90,
Chambers, J.R., Windschitl, P.D. Biases in comparative judgments:  e role of nonmotivated factors in
above-average and comparative-optimism eff ects. Psychological Bulletin, 130, 813–838.
Alicke_RT19523_C05.indd 104Alicke_RT19523_C05.indd 104 7/11/2005 11:06:53 AM7/11/2005 11:06:53 AM
Chambers, J. R., Windschitl, P. D., & Suls, J. (2003). Egocentrism, event frequency, and comparative
optimism: When what happens frequently is “more likely to happen to me.” Personality and Social
Psychology Bulletin, 29, 1343–1356.
Codol, J. P. (1975). On the so-called “superior conformity of the self” behavior: Twenty experimental
investigations. European Journal of Social Psycholog y, 5, 457–501.
College Board. (1976–1977). Student descriptive questionnaire. Princeton, NJ: Educational Testing
Cross, P. (1977). Not can but will college teachers be improved? New Directions for Higher Education, 17,
Dunning, D., & Cohen, G.L. (1992). Egocentric defi nitions of traits and abilities in social judgment.
Journal of Personality and Social Psycholog y, 63, 341–355.
Dunning, D., & Hayes, A. F. (1996). Evidence for egocentric comparison in social judgment. Journal of
Personality and Social Psychology, 71, 213–229.
Dunning, D., Meyerowitz, J. A., & Holzberg, A. D. (1989). Ambiguity and self-evaluation:  e role
of idiosyncratic trait defi nitions in self-serving assessments of ability. Journal of Personality and Social
Psychology, 57, 1082–1090.
Epley, N., & Dunning, D. (2000). Feeling “holier than though”: Are self-serving assessments produced by
errors in self or social prediction? Journal of Personality and Social Psychology, 79, 861–875.
Eiser, J. R., Pahl, S., & Prins, Y. R. A. (2001). Optimism, pessimism, and the direction of self-other com-
parisons. Journal of Experimental Social Psychology, 37, 77–84.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117–140.
Giladi, E. G., & Klar, Y. (2002). When standards are wide of the mark: Nonselective superiority and
inferiority biases in comparative judgments of objects and concepts. Journal of Experimental Psychology:
General, 131, 538–551.
Karylowski, J. J. (1990). Social reference points and accessibility of trait-related information in self-other
similarity judgments. Journal of Personality and Social Psycholog y, 58, 975–983.
Karylowski, J. J., & Skarzynska, K. (1992). Asymmetric self-other similarity judgments depend on priming
self-knowledge. Social Cognition, 10, 235–254.
Klar, Y. (2002). Way beyond compare: Nonselective superiority and inferiority biases in judging randomly
assigned group members relative to their peers. Journal of Experimental Social Psychology, 38, 331–351.
Klar, Y., & Giladi, E. E. (1997). No one in my group can be below the group’s average: A robust positivity
bias in favor of anonymous peers. Journal of Personality and Social Psychology, 73, 885–901.
Klein, S. B., & Loftus, J. (1993). The mental representation of trait and autobiographical kno wledge
about the self. In T. K. Srull & R. S. Wyer (Eds.), Advances in social cognition (Vol. 5, pp. 1–49). Hillsdale,
NJ: Erlbaum.
Klein, S. B., Loftus, J., & Burton, H. A. (1989). Two self-reference eff ects:  e importance of distinguishing
between self-descriptiveness judgments and autobiographical retrieval in self-referent encoding. Journal
of Personality and Social Psychology, 56, 853–865.
Klein, S. B., Loftus, J., Trafton, J. G., & Fuhrman, R. W. (1992). Use of exemplars and abstractions in
trait judgments: A model of trait knowledge about the self and others. Journal of Personality and Social
Psychology, 63, 739–753.
Kruger, J. (1999). Lake Woebegon be gone!  e “below-average eff ect” and the egocentric nature of com-
parative ability judgments. Journal of Personality and Social Psychology, 77, 221–232,
Matlin, M., & Stang, D. (1978). e Pollyanna principle: Selectivity of language, memory, and thought.
Cambridge, MA: Schenkman.
Moore, D. A., & Kim, T.G. (2003). Myopic social prediction and the solo comparison eff ect. Journal of
Personality and Social Psychology, 85, 1121–1135.
Nisbett, R. E., & Ross, L. (1980). Human inference: Strategies and shortcomings of social judgment. Englewood
Cliff s, NJ: Prentice-Hall.
Otten, W., & Van der Pligt, J. (1966). Context eff ects in the measurement of comparative optimism in
probability judgments. Journal of Social and Clinical Psychology, 15, 80–101.
Alicke_RT19523_C05.indd 105Alicke_RT19523_C05.indd 105 7/11/2005 11:06:53 AM7/11/2005 11:06:53 AM
Perloff , L. S., & Fetzer, B. K. (1986). Self-other judgments and perceived vulnerability to victimization.
Journal of Personality and Social Psycholog y, 50, 502–510.
Preston, C. E., & Harris, S. (1965). Psychology of drivers in traffi c accidents. Journal of Applied Psychology,
49, 284–288.
Reeder, G. D., & Brewer, M. B. (1979). A schematic model of dispositional attribution in interpersonal
perception. Psychological Review, 86, 61–79.
Rothbart, M., & Park, B. (1986). On the confi rmability and disconfi rmability of trait concepts. Journal of
Personality and Social Psychology, 50, 131–142.
Schlenker, B. R. (1980). Impression management:  e self-concept, social identity, and interpersonal relations.
Monterey, CA: Brooks/Cole.
Schkade, D. A., & Kahneman, D. (1998). Does living in California make people happy? A focusing illusion
in judgments of life satisfaction. Psychological Science, 9, 340–346.
Sears, D. O. (1983).  e person-positivity bias. Journal of Personality and Social Psychology, 44, 233–250.
Sedikides, C., & Gregg, A.P. (2003). Portraits of the self. In M. A. Hogg & J. Cooper (Eds.), Sage handbook
of social psychology (pp. 110–138). London: Sage.
Sedikides, C., & Strube, M. J. (1997). Self-evaluation: To thine own self be good, to tine own self be sure,
to thine own self be true, and to thine own self be better. In M. P. Zanna (Ed.), Advances in experimental
social psychology (Vol 29, pp. 209–269). New York: Academic Press.
Steele, C. M. (1988).  e psychology of self-affi rmation: Sustaining the integrity of the self. In L. Berkowitz
(Ed.), Advances in experimental social psychology (Vol. 21, pp. 261–302). New York: Academic Press.
Suls, J., Lemos, K., & Stewart, H. L. (2002). Self-esteem, construal, and comparisons with the self, friends,
and peers. Journal of Personality and Social Psycholog y, 82, 252–261.
Svenson, O. (1981). Are we all less risky and more skillful than our fellow drivers? Acta Psychologica, 47,
Swann, W. B. Jr., Rentfrow, P. J., & Guinn, J. S. (2003). Self-verifi cation:  e search for coherence. In M.
R. Leary and J. P. Tangney (Eds.), Handbook of Self and Identity (pp. 367–383). New York: Guilford
Tabachnik, N., Crocker, J., & Alloy, L. B. (1983). Depression, social comparison, and the false-consensus
eff ect. Journal of Personality and Social Psychology, 45, 688–699.
Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental
health. Psychological Bulletin, 103, 193–210.
Taylor, S. E., Lerner, J. S., Sherman, D. K., Sage, R. M., & McDowell, N. K. (2003). Are self-enhancing
cognitions associated with healthy or unhealthy biological profi les? Journal of Personality and Social
Psychology, 85, 605–615.
Tesser, A. (2000). On the confl uence of self-esteem maintenance mechanisms. Personality and Social
Psychology Review, 4, 290–299.
Weinstein, N. D. (1983). Reducing unrealistic optimism about illness susceptibility. Health Psychology,
2, 11–20.
Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and
Social Psychology, 39, 806–820.
Weinstein, N. D. (1984). Why it won’t happen to me: Perceptions of risk factors and susceptibility. Health
Psychology, 3, 431–457.
Weinstein, N. D., & Lachendro, E. (1982). Egocentrism as a source of unrealistic optimism. Personality
and Social Psychology Bulletin, 8, 195–200.
Windschitl, P. D., Kruger, J., & Sims, E. N. (2003).  e infl uence of egocentrism and focalism on
people’s optimism in competition: When what aff ects us equally aff ects me more. Journal of Personality
and Social Psychology, 85, 389–408.
Zuckerman, M. (1979). Attribution of success and failure revisited, or:  e motivational bias is alive and
well in attribution theory. Journal of Personality, 47, 245–287.
Alicke_RT19523_C05.indd 106Alicke_RT19523_C05.indd 106 7/11/2005 11:06:54 AM7/11/2005 11:06:54 AM
... The better-than-average effect describes people's tendency to evaluate their abilities, attributes, and personality traits as better than an average peer (Alicke & Govorun, 2005). This effect comes in line with the selfenhancement bias as self-insight (Allport, 1960;Kwan et al., 2004), where individuals perceive themselves in a better light than others do. ...
... Interestingly, the better-than-average effect feedback (Alicke & Govorun, 2005) was also assessed by our participants as more accurate than the real feedback, with no differences between the two types of false feedback. These results may offer a more clear direction toward unraveling the mechanism for these biases. ...
Full-text available
Despite personality measurement and feedback being pervasive practices, there are self-judgment biases that may impair their usage. We set out to analyze the differences between two kinds of false feedback and real feedback on personality regarding perceived accuracy and preference. We propose that there would be no differences between false and real feedback regarding perceived accuracy, but we expect differences regarding feedback preference. A sample of 146 students completed the IPIP-50 instrument that measured the Big 5 Factors and received three kinds of feedback - a general one (Barnum effect as false feedback), a positive one (Better-than-average effect as false feedback), and a real one. They rated each regarding accuracy and preference. Results indicate differences regarding both dependent variables. Participants perceive false feedback as more accurate than the real one. Moreover, they prefer positive feedback over the other two, and general feedback compared to the real one. We discuss both theoretical and practical implications, alongside a series of limitations and future research directions.
... Bias blind spot (e.g., Pronin et al., 2002a) Being convinced that mainly others succumb to biased information processing Hostile media bias (e.g., Vallone et al., 1985) Partisans perceiving media reports as biased toward the other side I am good. Better-than-average effect (e.g., Alicke & Govorun, 2005) Overestimating one's performance in relation to the performance of others Self-serving bias (e.g., Mullen & Riordan, 1988) Attributing one's failures externally but one's successes internally ...
Full-text available
One of the essential insights from psychological research is that people's information processing is often biased. By now, a number of different biases have been identified and empirically demonstrated. Unfortunately, however, these biases have often been examined in separate lines of research, thereby precluding the recognition of shared principles. Here we argue that several-so far mostly unrelated-biases (e.g., bias blind spot, hostile media bias, egocentric/ethnocentric bias, outcome bias) can be traced back to the combination of a fundamental prior belief and humans' tendency toward belief-consistent information processing. What varies between different biases is essentially the specific belief that guides information processing. More importantly, we propose that different biases even share the same underlying belief and differ only in the specific outcome of information processing that is assessed (i.e., the dependent variable), thus tapping into different manifestations of the same latent information processing. In other words, we propose for discussion a model that suffices to explain several different biases. We thereby suggest a more parsimonious approach compared with current theoretical explanations of these biases. We also generate novel hypotheses that follow directly from the integrative nature of our perspective.
... It is well-documented that people hold overly positive self-evaluations (Ferris, Johnson, & Sedikides, 2018;Taylor & Brown, 1988;Zell, Strickhouser, Sedikides, & Alicke, 2020). This widespread phenomenon where people overestimate their positive qualities and abilities and underestimate their negative qualities vis-a-vis others is known by many names including the "better than average effect" (Alicke & Govorun, 2005), self-enhancement bias (Kramer, Newton, & Pommerenke, 1993), the "Lake Wobegon Effect" derived from a fictional U.S. town in a long-running radio show where "all the women are strong, all the men are good looking, and all the children are above average" (Betts et al., 2011, p. 734), and the "Dunning-Kruger effect" (Dunning, 2011;Kruger & Dunning 1999) where individuals occupying inferior positions in numerous performance domains are unaware of just how deficient their expertise is. Worse yet, Dunning's (2005) research revealed that it is the most deeply incompetent people who make the most inflated self-assessments. ...
Full-text available
Small losses, measurable incremental negative workplace behaviors that do not align with organizational and societal norms often trigger a dynamic, magnifying process that produces greater undesirable outcomes, which once started, can often result in a downward spiral. Quickly addressing small problems can prevent minor misconduct and wrongdoing from escalating into greater difficulties later that can appear unstoppable. Failure to tackle the small issues and big problems will often follow. Managers should celebrate small victories or wins but also need to address small losses. Just as small wins can result in significant organizational gains, small losses can result in business losses. Keywords: workplace behavior, workplace misconduct, deviation amplifying processes
... Such estimates are imprecise (noisy), and tend to regress towards some common default. If that default guess is optimistic, and there is much evidence that people in general rate themselves above average [13][14][15], then it will be a gross overestimate for an unskilled person, and closer to the truth for a more-skilled person. Under this noise-plus-bias account, the DKE is attributable to uncertain self-estimation in the context of performance differences between more-and less-skilled people, and a general tendency for people to rate themselves as better than average. ...
Full-text available
For many intellectual tasks, the people with the least skill overestimate themselves the most, a pattern popularly known as the Dunning–Kruger effect (DKE). The dominant account of this effect depends on the idea that assessing the quality of one's performance (metacognition) requires the same mental resources as task performance itself (cognition). Unskilled people are said to suffer a dual burden : they lack the cognitive resources to perform well, and this deprives them of metacognitive insight into their failings. In this Registered Report , we applied recently developed methods for the measurement of metacognition to a matrix reasoning task, to test the dual-burden account. Metacognitive sensitivity (information exploited by metacognition) tracked performance closely, so less information was exploited by the metacognitive judgements of poor performers; but metacognitive efficiency (quality of metacognitive processing itself) was unrelated to performance. Metacognitive bias (overall tendency towards high or low confidence) was positively associated with performance, so poor performers were appropriately less confident—not more confident—than good performers. Crucially, these metacognitive factors did not cause the DKE pattern, which was driven overwhelmingly by performance scores. These results refute the dual-burden account and suggest that the classic DKE is a statistical regression artefact that tells us nothing much about metacognition.
... These findings are unsurprising, given a general interest in conferring respect on winners as part of combat etiquette (Pham et al., 2017). Interest in conferring respect following victory could reflect the proclivity to act prosocially following success, which may be coupled with a belief of oneself as a more gracious winner than the average person (Aknin et al., 2018;Alicke & Govorun, 2005). ...
Full-text available
Physical conflict has been historically prevalent throughout human evolution, with physically strong men possessing an advantage. To reduce the likelihood of incurring continued costs of conflict, opponents may engage in postconflict reconciliation to secure valuable social relationships. Two studies considered how formidability of male combatants informs expectations of reconciliatory behavior. In Study 1, participants reported expectations of respect exchanges between combatants, both following wins and losses, who were physically strong and weak. Study 2 tasked men with reporting their expectations for respect exchanges with strong and weak opponents following wins and losses. Strong targets were consistently expected to receive more respect following conflict. Nonetheless, male perceivers intended to display more respect against strong opponents regardless of fight outcome. Men’s upper body strength provides an important cue in shaping alliances for men, particularly when the potential costs of continued conflict are salient.
... We examined overrating (Gebauer et al., 2012), better-than-average (Alicke & Govorun, 2005), and overclaiming (Paulhus et al., 2003) effects in agentic and communal domains to test our predictions. Lastly, we examined the predictive utility of our newly proposed sanctity and heroism traits in explaining functioning in interpersonal relations and prosocialness as the most relevant communal aspects of human life. ...
In the current paper we introduce a new conceptualization of communal narcissism, as encompassed by two narcissistic strategies congruent with self-promotion and self-defensive motives. In addition, we posit that communal narcissistic strategies should be aligned with communal self-enhancement. Therefore, we propose narcissistic sanctity as an ego-boosting strategy and narcissistic heroism as an ego-defensive strategy. In a series of eleven studies (N = 5,606) we develop, validate, and employ what we will call the Narcissistic Sanctity and Heroism Concept. We found the scale to be a robust measure of communal narcissism and that the two postulated strategies are psychologically distinct while psychometrically sound. Specifically, we found that narcissistic sanctity was related to implicit and explicit communion, communal (but not agentic) self-enhancement and explained more socially (as compared to heroism) acceptable functioning in close relationships, being positively correlated with prosocialness via denying one’s egoistic motivations. Narcissistic heroism, on the other hand, was related both to agency and communion, it was unrelated to communal self-enhancement, explained less socially acceptable (as compared to sanctity) functioning in close relationships, and was related to less prosocialness via egoistic motivation. In sum, a newly proposed model of communal narcissism sheds new light on prior research on communal narcissism, explaining null relationships between communal self-presentation and actual behaviors in the communal domain.
This article considers the phenomenon of overconfidence whereby an individual, group or organization believes that it has more knowledge or skill in a particular domain than it actually possesses. It outlines the three distinct forms of confidence that have been identified in the literature: misestimation, misplacement and misprecision. It goes on to discuss various ways in which organizations can adapt their judgement processes to reduce the incidence of overconfidence, highlighting some real-world case studies. It ends with some observations and suggestions for future research in this complex area.
Competing successfully often means choosing the right competitions to enter, preparing for them adequately, and knowing when to quit. Many psychological phenomena, however, lead people to misjudge how their skills, talents, and prospects compare to peers they are competing against. Typically, these psychological tendencies lead people to a sense of illusory superiority, in which people overestimate their chances in competition. Example phenomena include relying on self-flattering conceptions of skill, focusing on optimistic scenarios while neglecting pessimistic ones, emphasizing one’s intentions and agency in producing outcomes while neglecting the impact of external forces, failing to have the competence to fully recognize competence in others, and quizzically not considering the skills of others when making decisions about whether to compete with them. Some specific circumstances, however, lead to underestimation of self. Those with negative self-views may discount their performances unduly. Top performers may not know how distinctive their skills are. In short, mistaken or biased views of the self may lead people to make decisions about competition that undermine their potential success in them.
This chapter provides an overview of theory and research on competition pools, namely, salient groups in which social comparisons occur routinely and exert a significant impact on self-evaluations. The authors conceptualize competition pools broadly to include competitions among friends, family, coworkers, teammates, and classmates. After a brief overview of social comparison theory, the authors describe the impact of social comparisons that arise in competition pools. Further, the authors note that effects of social comparisons often occur alongside effects of intrapersonal or temporal comparisons. Next, the authors discuss research on how people evaluate referents who outperform them in competition pools. This research, including work on the genius effect and referent status neglect, suggests that people selectively consider the skill level of the referent in order to maintain a positive self-image. The authors then turn to the distinction between local versus global competition pools. Informed by research on the big-fish-little-pond effect and local dominance, the authors argue that social comparisons are most impactful when made locally, among a few immediate others, as opposed to globally, with larger groups. Finally, the chapter concludes by highlighting broader implications of research on competition pools for social comparison theory as well as directions for future research.
Full-text available
Past research suggests that people believe that they perform socially desirable behaviors more frequently and socially undesirable behaviors less frequently than others (Goethals, 1986; Messick, Bloom, Boldizar, & Samuelson, 1985). The present research examined whether this perception also characterizes people's thinking about intelligent and unintelligent behaviors. In Study 1, subjects wrote lists of behaviors that they or others did. Subjects indicated that they performed more good and intelligent behaviors and fewer bad and unintelligent behaviors than others, although the magnitude of these differences was greater for good and bad acts than for intelligent and unintelligent ones. In Study 2, a different group of subjects judged the frequency with which the behaviors generated in the first study occur. While self-ascribed good behaviors were rated as occurring more frequently than the good acts of others, self-ascribed intelligent behaviors were not judged as more frequent than the intelligent acts of...
Full-text available
164 undergraduates rated the degree to which various traits represented desirable characteristics and the degree to which it was possible for a person to exert control over each of these characteristics. From these initial ratings, 154 trait adjectives for which 4 levels of desirability were crossed with 2 levels of controllability were selected. 88 undergraduates then rated the degree to which each of these traits characterized the self and the average college student. Results support the prediction that self-ratings in relation to average college student ratings would be increasingly positive as traits increased in desirability and that in conditions of high desirability, self-ratings in relation to average college student ratings would be greater for high- than for low-controllable traits, whereas in conditions of low desirability the opposite would occur. Results are discussed in terms of the adaptive advantages of maintaining a global self-concept that implies that positive characteristics are under personal control and that negative characteristics are caused by factors outside of personal control. Mean preratings of desirability and controllability are appended. (29 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Full-text available
Examined the role of contextual information such as comparison standard on self-other probability judgments regarding the occurrence of negative life events, which tend to be characterized by optimism. In Study 1, 80 undergraduates (mean age 19.5 yrs) completed a questionnaire on preventive behaviors and intentions. Ss judged 10 negative events while the comparison procedure, number of categories, and order of judgment were varied. In Study 2, 488 Ss (mean age 21.3 yrs) completed 2 questionnaires on the same 10 events, with a 1-wk interval between the 2. Preceding judgment, comparison procedures and time-lag were varied. Estimation using other-as-standard led to more optimism than using self-as-standard. Comparing self to others also led to more optimism than the reverse. Results show that transient reference points may reduce the tendency to see the self as better than others."
Three investigations are reported that examined the relation between self-appraisals and appraisals of others. In Experiment 1, subjects rated a series of valenced trait adjectives according to how well the traits described the self and others. Individuals displayed a pronounced “self-other bias,” such that positive attributes were rated as more descriptive of self than of others, whereas negative attributes were rated as less descriptive of self than of others. Furthermore, in contrast to C. R. Rogers's (1951) assertion that high self-esteem is associated with a comparable regard for others, the tendency for individuals to evaluate the self in more favorable terms than they evaluated people in general was particularly pronounced among those with high self-esteem. These findings were replicated and extended in Experiment 2, where it also was found that self-evaluations were more favorable than were evaluations of a friend and that individuals with high self-esteem were most likely to appraise their friend...
The asymmetry effect in self-other similarity judgments refers to a tendency for similarity judgments to be higher when self is used as a reference point in the comparison (“How similar is (Person X) to you?” questions) than when the self is used as a subject (“How similar are you to (Person X)?” questions). Two experiments were conducted to examine the impact of priming of self-knowledge on the asymmetry effect. As predicted from the Tversky's (1977) contrast model, there was a significant interaction between priming of self-knowledge and the direction of the comparison. Higher similarity judgments for “How similar is (Person X) to you?” questions as compared to “How similar are you to (Person X)?” questions were found in the Self-Primed condition but not in the Control condition or in the Other-Primed condition. Theoretical implications in terms of the organization of self-knowledge in memory and in terms of general processes underlying asymmetric similarity judgments are discussed.
Self-affirmation processes are being activated by information that threatens the perceived adequacy or integrity of the self and as running their course until this perception is restored through explanation, rationalization, and/or action. The purpose of these constant explanations (and rationalizations) is to maintain a phenomenal experience of the self-self-conceptions and images as adaptively and morally adequate—that is, as competent, good, coherent, unitary, stable, capable of free choice, capable of controlling important outcomes, and so on. The research reported in this chapter focuses on the way people cope with the implications of threat to their self-regard rather than on the way they cope with the threat itself. This chapter analyzes the way coping processes restore self-regard rather than the way they address the provoking threat itself.
Large samples of students in the Midwest and in Southern California rated satisfaction with life overall as well as with various aspects of life, for either themselves or someone similar to themselves in one of the two regions. Self-reported overall life satisfaction was the same in both regions, but participants who rated a similar other expected Californians to be more satisfied than Midwesterners. Climate-related aspects were rated as more important for someone living in another region than for someone in one's own region. Mediation analyses showed that satisfaction with climate and with cultural opportunities accounted for the higher overall life satisfaction predicted for Californians. Judgments of life satisfaction in a different location are susceptible to a focusing illusion: Easily observed and distinctive differences between locations are given more weight in such judgments than they will have in reality.